You can also see it recently for Mexico, nominal and real wages are rising while the share of _formal_ employement does as well, however there's a fall in productivity. Since Mexico has a relatively similar labor market to a 20th century European economy, this could be sign of a more general phenomenon under a soewhat specific set of policies, especially on labor and social security in the face of industrialization.
So, I don't have a story here. But what was driving changes in labor force participation during these periods? I'm guessing women entering the labor force, particularly for the 1979-90 period. But why would that drive the correlations in Europe but not the US?
I think that's a coherent story for that period (but I'd have to look at the gender breakdown to be sure). I was more focused on the later period when growth was slowing down, and at that point I'd *guess* that it was less women entering the labor force, and more about policy changes to encourage employment.
I have two quick thoughts, though their relevance depends on facts I am uncertain about.
1) It could just be a matter of methodology.
The description seems to imply that the European productivity residuals here are based on human capital measures for the whole population, whereas many of the most widely used U.S. productivity residuals (e.g. BLS-MFP or John Fernald's series) are based on measures of employed human capital and directly incorporate gender into their measurement of human capital. The Roy-style selection story makes a lot of sense to me when considering residuals based on population-level human capital measures, but methodologies that base their measurement of human capital/labor quality on employed workers would mute the effect of selection on measured productivity—that is generally the intended goal in using such approaches, though they probably are not 100% successful (e.g. selection still has an effect on imputed wages, which probably still filters through to measured productivity). So the difference could be rationalized as an apples-to-oranges comparison, though I am not sure whether this is relevant based on the description given here.
2. Even if the methodologies are consistent, the effect of women entering the labor force on measured productivity could differ considerably.
My impression is that historically the U.S. has been ahead of Europe when it comes to women's education, at least in terms of raw quantity (average years, high school/college completion, etc) if not quality (accessibility of different fields, ability to actually find work) or equity (equal pay, racial disparities). For example, women already accounted for a majority of college enrollment by the early/mid-1970s in the U.S. but were still a minority until the 1990s in Italy. To the extent that my impression is correct, I suspect that any gender gap in human capital was considerably smaller for the U.S. (at least in terms of the observables used in measures of productivity) and that college education was a better proxy for women's human capital in the U.S. (not having a college education is not very informative when very few women go to college). So the difference in educational context seems like a plausible potential account, though I certainly don't know enough to have any confidence in asserting that it actually drives the difference.
Love the description of productivity growth as a trash can containing all measurements
You can also see it recently for Mexico, nominal and real wages are rising while the share of _formal_ employement does as well, however there's a fall in productivity. Since Mexico has a relatively similar labor market to a 20th century European economy, this could be sign of a more general phenomenon under a soewhat specific set of policies, especially on labor and social security in the face of industrialization.
In France, it is argued that "apprenticeship" could explain 1/5 to 1/3 of the recent productivity slowdown (https://dares.travail-emploi.gouv.fr/sites/default/files/13a185dc0db8392304d42d8fd5948f41/Dares_Fovus5_Impact-hausse-alternance-sur-productivite.pdf )
So, I don't have a story here. But what was driving changes in labor force participation during these periods? I'm guessing women entering the labor force, particularly for the 1979-90 period. But why would that drive the correlations in Europe but not the US?
Like I said, I don't have a story.
I think that's a coherent story for that period (but I'd have to look at the gender breakdown to be sure). I was more focused on the later period when growth was slowing down, and at that point I'd *guess* that it was less women entering the labor force, and more about policy changes to encourage employment.
I have two quick thoughts, though their relevance depends on facts I am uncertain about.
1) It could just be a matter of methodology.
The description seems to imply that the European productivity residuals here are based on human capital measures for the whole population, whereas many of the most widely used U.S. productivity residuals (e.g. BLS-MFP or John Fernald's series) are based on measures of employed human capital and directly incorporate gender into their measurement of human capital. The Roy-style selection story makes a lot of sense to me when considering residuals based on population-level human capital measures, but methodologies that base their measurement of human capital/labor quality on employed workers would mute the effect of selection on measured productivity—that is generally the intended goal in using such approaches, though they probably are not 100% successful (e.g. selection still has an effect on imputed wages, which probably still filters through to measured productivity). So the difference could be rationalized as an apples-to-oranges comparison, though I am not sure whether this is relevant based on the description given here.
2. Even if the methodologies are consistent, the effect of women entering the labor force on measured productivity could differ considerably.
My impression is that historically the U.S. has been ahead of Europe when it comes to women's education, at least in terms of raw quantity (average years, high school/college completion, etc) if not quality (accessibility of different fields, ability to actually find work) or equity (equal pay, racial disparities). For example, women already accounted for a majority of college enrollment by the early/mid-1970s in the U.S. but were still a minority until the 1990s in Italy. To the extent that my impression is correct, I suspect that any gender gap in human capital was considerably smaller for the U.S. (at least in terms of the observables used in measures of productivity) and that college education was a better proxy for women's human capital in the U.S. (not having a college education is not very informative when very few women go to college). So the difference in educational context seems like a plausible potential account, though I certainly don't know enough to have any confidence in asserting that it actually drives the difference.